Modeling-to-generate alternatives with oemof
This repo hosts an implementation of a modeling-to-generate-alternatives (MGA) algorithm with some examples within the modeling framework oemof.
The idea of MGA is to investigate alternative solutions of an optimization model which can be different in other quantities but have nearly the same total system cost as the global optimum.
A well-known approach is the "hop, skip and jump" algorithm ([Brill1982]), which finds alternative solutions in the following way: After finding the global optimum of a LP, a new constraint is added which sets an upper bound on the total system costs such that they are smaller than the optimal value plus some tolerance. Then, the original objective function is dropped and instead, the euclidean distance to the previous solution is maximized. This is repeated several times to sample alternative solutions from the near-optimal feasible space.
The algorithm implemented here takes a different approach which is similar to the one described in [Neumann2021].
Generic functions implementing the main steps of the algorithm (setting an upper
bound on the cost, replacing the objective function of the model) are to be found
in mga.py.
How to use
Install the requirements listed in requirements.txt in your environment by calling:
pip install -r requirements.txt
Then, run python invest_optimize_all_technologies.py as an example application.
Results
Todo: Show plot.
References
| [Brill1982] | Brill, E. Downey, Shoou-Yuh Chang, and Lewis D. Hopkins. "Modeling to Generate Alternatives: The HSJ Approach and an Illustration Using a Problem in Land Use Planning." Management Science 28, no. 3 (1982): 221-35. Accessed July 6, 2021. http://www.jstor.org/stable/2630877. |
| [Neumann2021] | Neumann, Fabian, and Tom Brown. “The Near-Optimal Feasible Space of a Renewable Power System Model. ”Electric Power Systems Research 190 (January 2021): 106690. https://doi.org/10.1016/j.epsr.2020.106690. |